Xiang on Nostr: My second post - I toyed with ChatGPT asking some pedantic questions related to city ...
My second post - I toyed with ChatGPT asking some pedantic questions related to city network, e.g., on why saskia sassen grounded her world city theory largely, if not solely on presence of advanced producer service, and gained really professional answers.
What is so interesting behind chatgpt alike AI models which r all the rage currently is that they r large models built on corpus and other data at scales of billions.
But think of that: that's not how real intelligence is gained by human beings. Unlike models which feeds on billions of unicorn images to finally recoginse a unicorn, kids draw on very limited data and learn through concluding, and need to be shown only once before shouting out "UNICORN" next time they c a unicorn on screen.
Time to engage in a real lengthy talk on "what intelligence really is"?
What is so interesting behind chatgpt alike AI models which r all the rage currently is that they r large models built on corpus and other data at scales of billions.
But think of that: that's not how real intelligence is gained by human beings. Unlike models which feeds on billions of unicorn images to finally recoginse a unicorn, kids draw on very limited data and learn through concluding, and need to be shown only once before shouting out "UNICORN" next time they c a unicorn on screen.
Time to engage in a real lengthy talk on "what intelligence really is"?